Hybrid Cloud Workload Placement Strategy: Cloud, Colocation, and Considerations in Financial Modeling
While enterprise data center spend has been flat, and the number of physical data centers (mostly smaller) declining, cloud has enjoyed steady double-digit growth. Of late, some have suggested that movement of existing workload to the public cloud will be accelerated by a general rethinking of on-premises work requirements: firms will look even harder to get away from physical data centers.
As industry-recognized experts in what Gartner calls “Hybrid Cloud Workload Placement Strategy,” GTSG offers these thoughts on the decision process to determine “what runs where.” The issues may be particularly current, but nearly all of this was originally written over the past few years.
- “Begin with the End in Mind.”
Broadly, firms go to the cloud for some combination of
- Digital business enablement: for agility or developer productivity; or to drive innovation or utilize capabilities internal IT cannot deliver;
- “Next-generation” outsourcing: to gain flexible, scalable, improved infrastructure, or expand into new geographies; to create efficiencies or (in spots) manage cost.
It is essential to success to understand, measure, and communicate the goals of the initiative. Anything less rigorous leads to confusion and disappointment; and in this climate return from investment will be under exceptional scrutiny.
- Understand the financial case. According to Gartner, “Going to the cloud to save money” ranks barely above “the boss said so” a ranking of rationales to go to the cloud. If the goal is to “get out of the data center” – there are other potentially much quicker and much less expensive ways of accomplishing that.
- Cloud transition looks very different from a traditional outsourcing.
Many executives know the outsourcing business case: a straightforward enough comparison between “BAU” costs and the new (usually lower) sum of outsourced plus retained. With a cloud transition, we need also consider:
- “Top-of-the-business” consensus is required where transition will drive change from Capex to Opex
- Management systems and governance must be established, requiring investment in tooling, skills and training
- Cloud transitions are typically far longer than a traditional outsource transition period, and unlike the outsourcing scenario, no responsibility is held by the provider
- Transition timelines and costs can exceed internally budgeted expectations
- More workload may be retained on-prem (or in colo) than was anticipated –if there is no hard justification for migration; there is a cost to ready an application for the cloud
- Reduction in current state costs only occurs when discrete cost elements can be eliminated or demonstrably reduced. Examples include network connections (net against new connections required); software licenses; hardware; or personnel leaving the unit or measurably redeployed internal floor space (when it carries a value) or a colocation provider contract. In a typical cloud transition, this takes time.
Colocation as alternative. There are standard reasons to move to colo, whether simply looking to exit a non-core function including a better use of capex dollars, including the leveraging of connections and cloud onramps.
If a goal is to avoid reliance on workers in a physical data center, many colocation providers accept that responsibility via offer “remote hands” services.
Perhaps most importantly in a challenged economy, movement to colocation doesn’t require an application remediation effort.
* * * * *
The answer for any client always lies in the detailed analysis. Write us at Partners@GTSG.com. We are happy to talk further.